MusicFind: An Experiment In Music Discovery
David Schurman (my friend here at Brown and fellow STEAM leader) and I were both frustrated with the lack of visual music discovery tools out there. After searching for a virtual ‘music map’ of sorts, we were disappointed with the results. Most maps either only went one relational level deep or — even worse — were text based!
We realized that, in order for a related artists graph to be meaningful, two layers of depth are necessary, and relations between artists that are not the center artist have to be represented. This quite literally ties artists who are similar to each other together in the visualization.
This is particularly useful for seeing what different genres artists draw from. For instance, if you search “BØRNS” on Musicfind, you’ll see two groups form. One has artists from the new wave (well, one definition of new wave) scene, like Hippo Campus, Bleachers and WALK THE MOON. Another is populated with female-dominant indie pop groups, like Banks, Broods, and Wet. These visually distinct areas of the graph represent two different facets of BØRNS’ musical style.
As much as we’d love for this to be a fully functioning music discovery tool, it is still just an experiment. Cross-platform support is a goal for the future — it currently only works on Chrome due to some optimization requirements involving moving around and styling images in SVG. It would also be wonderful to give users full access to their Spotify accounts from the webpage so they can manage playlists while searching around the map. While we have abandoned work on Musicfind for now, we hope that sometime before the end of the year we can come back to it and turn it into the beautiful visual artist discovery tool that music enthusiasts deserve.
This project lives on the web! You can find it here.
Want to see the code? Shoot me a message and I’ll give you access to the GitHub.
